Introduction

(Description of questions that are being investigated)

There is a large number of call centres in Ontario and many more nationwide across Canada. Other than call centres itself, there are multiple companies and businesses using calling services as an essential form of communication for all of their clients and customers, whether it’s to help carry a certain service on call or just to gain feedback for customer service purposes. Findhelp Information Services is a charitable, nonprofit agency providing public access to information about community and social services through help lines, web directories, and specialized tools and training to help people find help.

We are Connected Consulting Corporation and we are a consulting firm specializing in improving a calling service to better a client/customer’s experience on call and make sure that each individual is being delivered the essential service perfectly in a tailored, timely manner. One of our biggest clients approaching us at this time is FindHelp. The 211 line is one the most called lines used on a daily basis by multiple individuals in Toronto. We are here to help FindHelp connect people and make sure that all parties are having a successful and efficient experience on both ends of the call. We have a team of three: Tashrif, Aaditya and Preet, working very closely with our client company to see what further steps we can take to find an improvised solution. We are utilizing the 211 caller data from the previous two years (2017-2019) and studying other datasets available in Ontario and additional secondary sources to present the data in an organized manner and set a finalized plan for FindHelp.

Research Questions

We have a set of questions to make our main focus while studying the datasets:

  1. What is the variance of different Level-1 services that receive high-volume calls?
  2. Which areas in Toronto most frequently require a call for services? Which areas are areas of high need? What can we do to improve the experiences of these high need areas? When are most people calling?
  3. Which services are most requested depending on area
  4. What are the average wait-times for calls for all services as a user must complete all 4 levels of categories?

DESCRIPTION OF OUR RESEARCH QUESTIONS:

  1. To begin with, we will try to see what Level-1 service is the most called service. We also want to see whether there is any variance between the immediate-attention services and informational services. Our biggest goal is to figure out what FindHelp clients are calling for the most and which service is an essential need on a daily basis.

  2. We want more information about the caller’s geographical characteristics. We are mainly focusing on the location portion for the research question. If we can locate hot-spots for the incoming calls on the 211 line using the postal-code in Toronto, then we can move on to figuring out how to improve and set a system to better the calling experience for all users of the area.

  3. Something similar to question (2), we want to further research based on the geographics once again and see what services are highly used in a particular area and set a system for those certain services to make a more transparent communication service for both ends of the 211 line.

  4. Lastly, we want to compare call-wait times. Call waiting is very annoying and lengthy sometimes. We want to take data from both the 211 caller data and other datasets to explore the regression throughout the years and see how we can find a solution to help decrease the call-wait times to improve the experience on a larger scale for the callers of FindHelp.

Data

To further study our final outlook on this assignment, along with the 211 Caller data, we have chosen a secondary dataset found from the Ontario Data Catalogue. We found a dataset computed by ServiceOntario which puts together data based on calls handled and the average wait time per category of calls. ServiceOntario is one of the largest kiosk services funded and run by the government to serve the citizens of Ontario with their needs such as; Health card registration, Birth, marriage and death certificates, and Driver and vehicle licensing. We truly believe that the dataset from ServiceOntario is an essential source to the success of attempting to create calling services at FindHelp better or flow more efficiently.

Analysis

  1. Initially we observed which Level-1 service received the most calls; and we found that Basic needs, Criminal Justice and Health care were amongst the most called services. After reviewing the data we found that more calls were directed towards informational services rather than emergency services.
  1. To tackle the question of which areas required the most service, we needed to analyze the number of calls that each area in Toronto was making to the client. After filtering the 211 Caller Data, of 200,000 calls, with only 30,000 calls having recorded postal area, we had to complete some more analysis of the data., After some analysis of the data and looking through over 100 areas in Toronto, we came to a conclusion. The top three areas that made the most calls were (by postal area): M5A, M1K, M4C. There are still over 100 more areas with many calls, however these areas are the ones with the largest total number of calls. A plot to the analysis can be seen below.

Or given more visually by a map.

What can we do to mitigate the number of calls from these high frequency areas? There must be certain services that these areas require. So we must complete a further analysis of the data. After some discovery work, there were a few more details we uncovered for the data. The top three areas all required the Basic Needs service the most. After some thought, in general although Basic Needs is not the most requested service for all areas, it competes usually for the highest spot, and a higher focus should be given to the Basic Needs calling services. Something such as more receivers, or shorter waiting times for basic needs should be implemented. Among other services, Community Services and Health Care were the other two contending services, and hence should also be given higher priority as there are higher callers in most communities for these services.

  1. Another method in which we could improve the 211 Caller Service, is that we could analyze which service is used most by any given area. This will allow our client to fine tune their services by area providing a better experience to callers from their area by providing services faster.

In our analysis of the 211 Caller Data, we must compute the number of calls per service by area. Then, we can plot this onto a map of Toronto.

Given this graph, we can analyze each area point. Given the outermost ring in the area, we will be able to see which service is used most. This will allow us to tune our results of the caller service by area.

We can also utilize the table above as a tool to determine which services are most required per area.

  1. Calling waiting is just one of the steps in the process of getting your call finally connected. The other steps include; selecting multiple options by dialing in numbers and also being put on hold (with instrumental music) before the call is even answered. This wait can be very lengthy and sometimes annoying for some clients. After studying very closely and looking for additional datasets, we came across the Service Ontario dataset. This dataset provides us the essential information for average wait times for their over-the-phone service over the previous four years (2016-2019).

The very first data representation that came to our mind was scatterplot. Since there are multiple calls made about multiple service categories within the same month, we decided to see the change in the average call-wait times year-by-year. Plot P51 shows the following relationship:

## data: service, month, year, calls_handled, average_wait, X, X.1, X.2,
##   X.3, X.4, X.5, X.6, X.7, X.8, X.9, X.10, X.11, X.12, X.13, X.14,
##   X.15, X.16, X.17, X.18, X.19, X.20, X.21, X.22, X.23, X.24, X.25,
##   X.26, X.27, X.28 [766x34]
## mapping:  x = ~year, y = ~average_wait
## faceting: <ggproto object: Class FacetNull, Facet, gg>
##     compute_layout: function
##     draw_back: function
##     draw_front: function
##     draw_labels: function
##     draw_panels: function
##     finish_data: function
##     init_scales: function
##     map_data: function
##     params: list
##     setup_data: function
##     setup_params: function
##     shrink: TRUE
##     train_scales: function
##     vars: function
##     super:  <ggproto object: Class FacetNull, Facet, gg>
## -----------------------------------
## geom_point: na.rm = FALSE
## stat_identity: na.rm = FALSE
## position_identity 
## 
## geom_smooth: na.rm = FALSE, se = TRUE
## stat_smooth: na.rm = FALSE, se = TRUE, method = auto, formula = y ~ x
## position_identity

This plot shows a scatter plot, where the dots are plotted based on the average minutes the call had lasted for the recorded call. The red line is the line-of-regression which shows the relationship of the x and y axis as a line plot on the same graph. The grey shaded region represents the 95% confidence interval (standard error) of the relationship. Based on plot P51, it is unanimously evident that the average call-time increases year-by-year.

## data: service, month, year, calls_handled, average_wait, X, X.1, X.2,
##   X.3, X.4, X.5, X.6, X.7, X.8, X.9, X.10, X.11, X.12, X.13, X.14,
##   X.15, X.16, X.17, X.18, X.19, X.20, X.21, X.22, X.23, X.24, X.25,
##   X.26, X.27, X.28 [766x34]
## mapping:  x = ~year, y = ~average_wait
## faceting: <ggproto object: Class FacetWrap, Facet, gg>
##     compute_layout: function
##     draw_back: function
##     draw_front: function
##     draw_labels: function
##     draw_panels: function
##     finish_data: function
##     init_scales: function
##     map_data: function
##     params: list
##     setup_data: function
##     setup_params: function
##     shrink: TRUE
##     train_scales: function
##     vars: function
##     super:  <ggproto object: Class FacetWrap, Facet, gg>
## -----------------------------------
## geom_point: na.rm = FALSE
## stat_identity: na.rm = FALSE
## position_identity

We can also see the same relation for each service category Service Ontario offers where the average call-wait time minutes increase by year: 2016<2017<2018<2019.

After taking an in-depth study of the Service Ontario dataset, we can see some observations and computations which provided us a better understanding of the data source. ServiceOntario receives an average call volume of 3,302,241 calls per year and they averaged a total of 1.08 (in minutes) of call wait times for all categories of calls combined (from Jan. 2016 to Sep. 2019). We thought that this was not enough data to create a justified solution to the problem or conclude any information for call waiting. This led us to carefully study the call wait times per year. As we computed the average per year, we noticed that the average in 2016 (0.78 wait time minutes) increased in 2019 (1.85 wait time minutes). However, the average call volume per year decreased. This dataset clearly presents the fact that clients are having to wait longer than usual for getting solutions for their service and questions.

Now, we know that FindHelp has 4 levels of categories to choose from. After making a test call as well, it is very evident that selecting the four levels and then being put on hold as well can become a lengthy process. The biggest change we would like to make to FindHelp’s service is to transform 4 levels into 3 levels. All the Level-4 categories/options will be regrouped into categorical groups that correlate with each other. In particular, all the immediate-attention services (ie. emergency) should be grouped under one category and individuals will only have to select directly from Level-1 categories. This will help FindHelp keep all immediate-attention services as express services and we have reduced the call-wait time for the other categories by making it only 3 levels.

Summary

After completing the research for the questions about the 211 Caller dataset, we have come to multiple conclusions and discovered ways to improve the FindHelp Information Service. In general, the wait times and queue times for basic needs service needs to be lowered, as it is by far the most utilized service in the level 1 category. Aswell, area specific experiences can be offered to improve service based on most used services in the area. It’s been observed that a large majority of calls to 211 have been made within close proximity of each other, and in certain core areas of downtown Toronto the density of those calls far exceed regions like Scarborough. Lastly, we will change the service to just 3 levels of options rather than 4 and also add an expedited menu service which benefits the immediate services the most. Therefore, these changes will help improve FindHelp and keep everyone connected on both ends of the 211 line!